Managing Uncertain Tax Liabilities from Sham Trusts in Financial Remedies: Reverse Contingent Lump Sums and Unequal Sharing of Matrimonial Debt – Commentary on Michael v Michael [2025] EWCA Civ 1668

Managing Uncertain Tax Liabilities from Sham Trusts in Financial Remedies: Reverse Contingent Lump Sums and Unequal Sharing of Matrimonial Debt
Commentary on Michael v Michael [2025] EWCA Civ 1668

1. Introduction

Michael v Michael [2025] EWCA Civ 1668 is an important Court of Appeal (Civil Division) decision in the field of financial remedies following divorce. Lord Justice Moylan (with whom Lady Justice King agreed) refused the husband’s application for permission to appeal a substantial financial remedy order made by HHJ Hess (sitting as a Deputy High Court Judge).

The case is especially significant for three reasons:

  • It confirms the court’s wide discretion to deal with large and highly uncertain tax liabilities (here, arising from a sham trust and a restructuring) by using reverse contingent lump sum orders and by capping a spouse’s contribution to such liabilities.
  • It endorses the approach that a trial judge need not select a single precise figure for an uncertain tax exposure where the evidence only supports a broad range; instead, the judge may use that range to design a fair mechanism for sharing risk.
  • It reinforces both the high threshold for appellate interference with findings on valuation and distribution in financial remedy cases, and the court’s flexible discretion to consider post-valuation events selectively when valuing business and property interests.

The parties were a husband and wife with total capital resources of approximately £40 million, most of it tied up in a complex corporate and (purported) trust structure. A central issue was the treatment of a potentially enormous, but highly uncertain, tax liability generated when the family trust (the “CN Trust”) was found to be a sham, with consequential UK tax implications on a 2020 restructuring.

The husband sought permission to appeal on a wide range of grounds, but on paper permission was only adjourned for oral hearing on three discrete grounds:

  1. Ground 1 – the judge’s treatment of the potential HMRC liability arising from the sham trust findings, including:
    • alleged failure to “grapple” with and quantify the likely tax liability;
    • the mechanism of a reverse contingent lump sum order; and
    • the imposition of a £7.5 million cap on the wife’s contribution.
  2. Ground 3(c) – the judge’s valuation of a development known as LR, held within the husband’s company structure.
  3. Ground 3(d) – the judge’s valuation of PEL Ltd, a wholly owned subsidiary of the husband’s UK holding company.

The Court of Appeal rejected each ground as having no real prospect of success and found no compelling reason to grant permission. In doing so, it provided a structured affirmation of several key principles of financial remedy adjudication and appellate restraint.

2. Factual and Procedural Background

2.1 The asset base and corporate structure

The parties’ capital resources were found by HHJ Hess to be about £40 million. A very substantial proportion of that wealth was held in a complex corporate and trust arrangement centred on:

  • CN Trust – a trust that, on its face, owned CNH Ltd and ultimately the business interests.
  • CNH Ltd – a Cypriot company formerly holding the group’s operating subsidiaries.
  • H (UK) Ltd (“H(UK)” or “H (UK)”) – an English holding company incorporated in November 2020, which came to own:
    • 100% of PEL Ltd (an English company with property interests);
    • 100% of HI Ltd (“HIL”), a Cypriot entity owed over £8.5 million within the group; and
    • 50% stakes and similar interests in a number of development and investment vehicles and an LLP.

2.2 Preliminary finding: the CN Trust as a sham

Following a 12-day preliminary issues hearing, HHJ Hess delivered a judgment on 12 August 2024 ([2024] EWFC 463), in which he held that the CN Trust was a sham. The effect was that:

  • The assets purportedly held by the trust – in particular 100% of H (UK) Ltd – were in fact beneficially owned by the husband.
  • Because of the 2020 restructuring, the court’s sham finding created a potentially very large UK tax liability for the husband:
    • On 24 November 2020, the UK group subsidiaries were moved from the Cypriot structure (HI Ltd, owned by CNH Ltd) into the newly incorporated H (UK) Ltd.
    • On the same day, CNH Ltd transferred its shares in H (UK) Ltd to the “trust” (i.e. in legal reality, to the husband) for nominal consideration of £1.
    • If the trust was a sham and the assets always belonged to the husband, HMRC might treat the share transfer as an income distribution of the full open market value of H (UK) Ltd (about £30.2 million), triggering income tax, interest and penalties.

This potential liability was analysed by a single joint tax expert, who modelled three scenarios. Depending on factual and discretionary elements (classification of the transaction; HMRC’s assessment of culpability and co-operation; whether disclosure was “prompted” or “unprompted”), the combined tax, interest and penalties could fall anywhere between roughly £6 million and £30 million, with an intermediate scenario around £18 million.

2.3 The financial remedy hearing and order

At the main financial remedy hearing, the court had the benefit of:

  • A single joint property valuation expert (valuations as at 31 March 2024).
  • A single joint chartered accountant (business valuations of the various companies and LLPs, again with an effective reference date of 31 March 2024).
  • The single joint tax expert dealing with the CN Trust-derived exposure.

HHJ Hess had to:

  • Value the husband’s interest in H (UK) Ltd, including several contentious developments (particularly WGR, CR, and LR);
  • Decide how to allocate
  • Address ancillary issues including “add-back” and valuation disputes over other companies (e.g. MBL) and debts; and
  • Determine an overall fair financial remedy outcome under the Matrimonial Causes Act 1973, s 25.

He ultimately concluded that:

  • Total capital resources were about £40 million.
  • The value of the husband’s shares in H (UK) Ltd should be taken at around £30.5 million, after adjustments to the expert’s figures and a minority discount.
  • The HMRC exposure could be anywhere between about £6 million and £30 million.
  • To deal with that uncertainty, he would use a reverse contingent lump sum mechanism so that the husband could seek reimbursement of up to 50% of the tax actually paid, but capped at £7.5 million.
  • The wife would receive a series of lump sums totalling £15 million (the final £7.5 million being due in September 2026), leaving the husband with largely illiquid, risk-laden business assets.

2.4 The proposed appeal

The husband applied for permission to appeal. On paper, Moylan LJ refused permission on most grounds but adjourned three grounds to an oral hearing:

  • Ground 1 – alleged errors in dealing with the sham trust tax liability:
    • failure to “attempt to quantify” the liability as best he could;
    • flaws in the mechanism requiring the husband to pay the wife first, then HMRC, then recover from the wife;
    • unjustified imposition of a £7.5 million cap on the wife’s contribution.
  • Ground 3(c) – error in valuing the LR development held through H (UK) Ltd.
  • Ground 3(d) – error in valuing PEL Ltd, a wholly-owned subsidiary of H (UK) Ltd.

At the Court of Appeal hearing, the husband was represented by Mr Wagstaffe KC and Mr Lewis; the wife by Mr Haggie, with a skeleton argument by Mr Pocock KC and Mr Haggie. All counsel had appeared below.

3. Summary of the Court of Appeal’s Decision

Lord Justice Moylan dismissed the application for permission to appeal on all remaining grounds. His core conclusions were:

  • The judge did “grapple” adequately with the tax liability:
    • HHJ Hess properly accepted the expert’s evidence that the liability could reasonably fall within a very wide range, from around £6 million to £30 million.
    • The evidence did not permit a more precise finding, and the judge was neither required nor entitled to engage in false precision by arbitrarily narrowing the range.
  • The reverse contingent lump sum with a £7.5 million cap was within the judge’s discretion:
    • The tax was indeed a matrimonial “debt”, but the judge was entitled to load a greater share onto the husband because the husband was the principal cause of the liability (through creating and defending the sham structure).
  • The structure requiring the husband to pay HMRC before recouping from the wife was not unfair:
    • The final lump sum to the wife is not due until September 2026, giving the husband enough time to address the tax position.
    • The mechanism appropriately conditions the wife’s obligation on the husband having actually paid HMRC.
  • The challenges to the valuations of LR and PEL were insubstantial and not appealable:
    • Any difference between the judge’s figure for H (UK) Ltd and the husband’s own suggested figure was immaterial in the context of a £40 million estate (about 4%).
    • The judge’s selective approach to post-valuation events (adjusting LR but not PEL) was squarely within his discretion.
    • There was no rationally insupportable finding, misdirection in law, or demonstrable failure to consider relevant evidence.
  • The proposed appeal raised no point with a real prospect of success, nor any compelling reason for permission:
    • The award’s effect, even in various tax-outcome scenarios, fell comfortably within the “generous ambit of discretion” afforded to a trial judge in financial remedies.
    • The husband’s attempt to introduce a new argument based on the Corporation Tax Act 2010 at the appellate stage was impermissible.

4. Detailed Analysis

4.1 The Potential HMRC Liability from the Sham Trust

4.1.1 Nature and scale of the tax exposure

The tax issue arose because the 2020 restructuring was re-characterised in light of the sham trust finding:

  • On 24 November 2020, the shares in H (UK) Ltd (said to be worth about £30.2 million) were transferred from CNH Ltd (a Cypriot company) to the CN Trust.
  • Once the “trust” was held to be a sham, the assets were seen as having always been beneficially owned by the husband.
  • HMRC might therefore treat the transaction as an income distribution to the husband of the full market value of H (UK) Ltd, giving rise to a UK income tax liability, due by 31 January 2022, plus interest and potentially up to 100% penalties depending on culpability.

The single joint tax expert modelled three broad scenarios, giving approximate total exposures (tax + interest + penalties) of £5.9 million, £17.98 million and £28.4 million. In evidence, he emphasised that:

  • There was great uncertainty as to how HMRC would characterise the transaction (capital distribution, employment income, unpaid consideration for a share purchase, etc.).
  • The ultimate outcome would depend heavily on the discretion of the HMRC caseworker(s), especially in relation to penalties and the assessment of culpability.
  • All the modelled outcomes remained possible on the facts.

The husband argued that, in oral evidence, the expert had effectively “jettisoned” the lowest figure, leaving a realistic range of roughly £20–35 million, and encouraged the judge to fix on a “best” estimate of £27.4 million. The wife, by contrast, invited the judge to adopt an approximate figure or range rather than be drawn into pseudo-precision when the underlying process was so inherently discretionary and fact-sensitive.

4.1.2 Was the judge obliged to choose a single figure?

The husband’s main complaint on appeal (Ground 1) was that the judge abdicated his duty by refusing to quantify the liability beyond the wide range of £6–30 million. Reliance was placed on Lewison LJ’s observation in Fage UK Ltd v Chobani UK Ltd [2014] EWCA Civ 5 at [115] that the primary function of a first-instance judge is to find facts and identify crucial legal points.

Moylan LJ rejected this characterisation. Analysing the judgment (paras 75–82), he held that:

  • The judge did grapple with the tax liability “as best he could” on the evidence.
  • He explicitly acknowledged that:
    • The potential outcomes formed a “very wide range”, from about £6 million (best case) to about £30 million (worst case).
    • It was impossible to predict with any confidence where within that band HMRC would land, given both factual uncertainties and the discretionary nature of interest/penalty decisions.
  • The judge’s conclusion was that the evidence did not permit any further refinement of the likely liability, either by eliminating the low-end scenario or by narrowing the range.

Moylan LJ emphasised two important points of principle here:

  1. Findings of fact are purposive: A trial judge’s findings exist to enable the court to carry out the statutory exercise fairly – here, to determine a fair financial remedy under s 25 MCA 1973. The level of specificity required is determined by the evidence and by what is needed to fashion a just solution. There is no freestanding duty to select a single figure if the evidence does not justify it.
  2. False precision is impermissible: It would have been wrong for the judge to narrow the range artificially simply because the parties pressed him to do so. To discard the lower scenario or fix on a central figure (like £27.4 million) without adequate evidential justification would itself have been vulnerable to appellate criticism.

On the husband’s suggestion that the lowest scenario should have been excluded because of the husband’s alleged divestment of CNH Ltd to an unrelated company (B Holdings), Moylan LJ noted:

  • This ownership issue had not been properly explored at trial, despite its alleged importance.
  • The husband’s case about B Holdings was internally inconsistent; for example, he had initially tried to deny that HIL (owed £8.5 million by H (UK) Ltd and PEL) was owned by H (UK) Ltd, claiming it remained with CNH Ltd/B Holdings, but that position was later abandoned.
  • The factual uncertainty surrounding CNH’s ownership was therefore not a secure foundation for excluding the lowest tax scenario.

Further, a new argument about the effect of the Corporation Tax Act 2010 was raised by the husband only on appeal. Citing the principle in Singh v Dass [2019] EWCA Civ 360, Moylan LJ held this was impermissible: such a point required expert evidence and ought to have been raised below.

Accordingly, the judge’s acceptance of a broad £6–30 million range was a classic evaluative finding of fact. Applying the appellate test derived from Henderson v Foxworth Investments Ltd, Re R (Children), and Lidl v Tesco, the Court of Appeal held that:

  • There was no material error of law;
  • No critical finding lacked an evidential basis;
  • No demonstrable misunderstanding or failure to consider relevant evidence was shown; and
  • The conclusion was plainly one the judge was entitled to reach.

4.1.3 The reverse contingent lump sum and the cap

Confronted with an uncertain but potentially ruinous tax liability, HHJ Hess adopted a formulaic solution. He ordered that:

  • The husband must keep the wife informed of any HMRC investigation or determination.
  • The wife must take all lawful reasonable steps within her power to minimise the tax (for example, through co-operation with disclosure, if relevant) and keep the husband informed.
  • If a liability does in fact arise and the husband actually pays the sums due to HMRC, he may recover from the wife:
    • 50% of the amount he has paid; but
    • subject to a cap of £7.5 million.
  • The legal mechanism is a reverse contingent lump sum—that is, a lump sum payable by the wife to the husband only if and when the specified contingency (actual payment of the HMRC liability) occurs.
  • There is no “long-stop” date: the reimbursement obligation endures as long as the husband’s potential liability to HMRC remains live.

The husband contended that the imposition of the £7.5 million cap was wrong in principle because the tax liability was a “matrimonial debt” which should be shared equally as part of the sharing principle derived from cases such as White v White and Miller; McFarlane (though these authorities are not expressly cited in the judgment). He argued that, whatever the final tax figure, the wife should bear 50% with no ceiling.

Moylan LJ accepted that HMRC liabilities arising during the marriage can be seen as matrimonial debts. However, he upheld the judge’s view that an unequal division of that debt was justified. Key points:

  • The trial judge had engaged in a careful conduct assessment:
    • The husband’s creation of the sham structure and his conduct in defending it made him the “principal cause” of the tax risk.
    • Although the wife had, during the marriage, acquiesced in the structure and benefited from it, the court could not ignore that \without her sham application she would never have accessed a fair share of the assets at all.
  • In light of s 25(2)(g) MCA 1973 (conduct), and the broad fairness-based discretion in financial remedies, the judge was entitled to decide that:
    • The wife should bear some, but not unlimited, exposure to the liability; and
    • The husband should shoulder the bulk of any tax at the top end of the range, as the party most responsible for its existence.
  • The judge expressly recognised that the cap might disadvantage the husband if the liability materialised towards the higher end (e.g. £30 million), but considered that this trade-off was nonetheless fair.

Moylan LJ concluded that this was a classic discretionary calibration, well within the judge’s remit, and that the cap was legitimately grounded in the judge’s assessment of responsibility and conduct.

4.1.4 The sequencing complaint: “pay now, recoup later”

The husband also argued that the mechanism unfairly required him:

  1. First, to pay the lump sums ordered to the wife;
  2. Second, to pay whatever HMRC demanded; and only thereafter
  3. To reclaim up to £7.5 million from the wife.

This, he said, imposed an unrealistic and oppressive burden, especially if HMRC imposed a high-end liability.

Moylan LJ considered this argument unpersuasive, noting that:

  • The last lump sum to the wife (of £7.5 million) was not due until 1 September 2026.
  • There was, as yet, no evidence that the tax issue could not realistically be dealt with by that time.
  • The conditionality of the wife’s liability on actual payment by the husband to HMRC added clarity and protected the wife from being dragged into disputes about theoretical or contested tax figures.

He also observed that, if HMRC’s liability had been clearly established before the last lump sum fell due, the enforcement court could take that into account when addressing any enforcement application concerning that sum. Overall, the structure was not shown to be unfair or irrational.

4.1.5 Overall evaluation of the tax treatment

At paragraph 42 of his judgment, Moylan LJ highlighted the distributional implications of the order under different tax outcomes:

  • If tax were about £6 million:
    • Husband: £22 million (65%)
    • Wife: £12 million (35%)
  • If tax were about £30 million:
    • Husband: £2.5 million (25%)
    • Wife: £7.5 million (75%)
  • On the husband’s own suggested figure of £27.4 million:
    • Husband: £5.1 million (40%)
    • Wife: £7.5 million (60%)

These possible outcomes remained within a band that the trial judge regarded as justified by the parties’ circumstances, the husband’s conduct, and the nature of the assets (the husband keeping illiquid, risk-laden business assets, versus the wife receiving liquid resources). The Court of Appeal agreed, stressing that the award was “well within the ambit of his discretion” and provided a “fair balance between the competing positions”.

4.2 Valuation of H (UK) Ltd: LR and PEL

4.2.1 General approach to valuations and post-valuation events

HHJ Hess summarised the law on valuing shares in private companies, stressing the need for caution in placing too much weight on expert valuations. He acknowledged the accepted principle that assets in financial remedy cases are ordinarily valued as at the trial date, but also that, pragmatically, expert valuation exercises tend to use an agreed “target date” (here 31 March 2024).

Crucially, he stated (in relation to the CR development, but in terms endorsed by the Court of Appeal) that:

  • The court has a discretion to look at “post valuation events” (i.e. changes after the expert date but before judgment) if it is fair to do so.
  • Equally, the court can decide not to take such events into account if doing so would be unfair—for example, if it would enable one party to “cherry pick” favourable developments while ignoring adverse movements.

This discretion was then exercised differently for different assets: the judge declined to revisit the valuation of CR but did increase the value of LR, while refusing to adjust the valuation of PEL. The husband attacked this as inconsistent. The Court of Appeal’s analysis clarifies that:

  • Selective adjustment is not inherently inconsistent or erroneous.
  • The question is whether the judge’s decision on each asset was open to him on the evidence and applied principled reasoning.

4.2.2 LR development (Ground 3(c))

The LR development had been valued by the single joint property expert on an “incomplete” basis:

  • Gross development value (GDV): £32.5 million;
  • Residual market value (as at March 2024, incomplete): £15.36 million;
  • Prospective annual rent used: £1.9 million.

The wife argued that this approach produced an artificially depressed value given that the development was very close to completion (around three months away), and that:

  • A separate bank valuation at the same time suggested a GDV of about £40 million with an annual rent of £1.9 million.
  • In reality, the development was later let at an annual rent of £2.75 million, implying an even higher capital value (an estimated £47 million on an equivalent yield.

The wife’s case was that the £15.36 million residual assumed an unreal scenario of selling an incomplete project at a discount to a third party who would “pick up the pieces”, whereas actual completion and letting rendered that pessimistic assumption obsolete. She accepted that any increase in value should be net of the actual borrowing and remaining build costs, which she had factored into her calculations.

HHJ Hess decided to:

  • Increase LR’s gross value by £10 million;
  • Make deductions “along the same lines” as WGR (including tax etc.), totalling £8 million;
  • Thus add a net £4 million to H (UK) Ltd’s value.

The husband argued on appeal that:

  • It was wrong in principle to depart from the single joint expert’s “incomplete” valuation, given that the agreed valuation date was March 2024.
  • If any uplift was taken to reflect completion and letting, the judge should have also taken full account of additional costs incurred in achieving that uplift, which he had allegedly failed to do.

Moylan LJ rejected this challenge:

  • The judge had a discretion to consider post-valuation events and plainly exercised it on a principled basis, taking into account the fact of completion and the higher rent.
  • The evidence about additional build costs was unclear; the husband’s late statement did not directly identify incremental costs since the expert valuation, and counsel struggled to articulate a coherent figure.
  • Even on the husband’s own figures, any additional build cost was comfortably covered by the “cushion” between:
    • The expert’s GDV of £32.5 million; and
    • The judge’s effective gross valuation of £25.3 million (original 15.3 + 10 uplift), still well below the GDV and below the bank/inferred valuations.
  • In other words, the uplift was conservative in context.

Importantly, the Court of Appeal also noted that the difference between the judge’s overall figure for H (UK) Ltd (£30.5 million net) and the husband’s own proposed figure (£28.9 million) was only about £1.6 million, which is 4% of the total estate. In a £40 million case, such a variation is not remotely sufficient to justify the expense and delay of an appeal—particularly when there is no clear error of law or irrationality.

4.2.3 PEL Ltd (Ground 3(d))

PEL was valued by the single joint expert at about £11.35 million, on the assumption that the company remained a going concern. The husband argued that its value had fallen since March 2024 owing to two developments:

  1. Increased borrowings:
    • Bank borrowing had risen from £4.1 million to £5 million by November 2024, and then to £6.8 million after a further loan of £1.8 million in late November 2024.
    • By the date of the final hearing, accrued interest had taken the total to £7.2 million.
  2. Receivership and enforcement:
    • Receivers had been appointed by the court to enforce arrears owed to the wife under:
      • Legal services payment orders;
      • Maintenance pending suit orders; and
      • An £850,000 payment on account of costs.
    • The receivers’ first report provided “low end” and “high end” valuations of £8.8 million and £11.4 million respectively, assuming potential auction sales.
    • The bank’s usual policy in receivership was to sell by public auction, which might achieve lower prices than private treaty sales.

The husband contended that:

  • There was now “a significant risk” that PEL’s properties would be sold at auction, realising less than the expert’s value.
  • The process of realising assets to pay the wife’s interim awards would crystallise corporation tax liabilities within PEL and further personal tax liabilities on distributions to the husband.
  • Accordingly, at least £2 million should be knocked off PEL’s value to reflect these costs, even though he suggested that the value might be nil.

HHJ Hess declined to adjust the expert’s valuation. He effectively treated PEL as still worth the expert figure, despite the receivership and additional borrowings. Moylan LJ upheld this approach for several reasons:

  • Nature of the receivers’ valuations:
    • The “low end” figure was explicitly framed as an auction value; the bank’s policy of preferring auction did not mean that auction sales were inevitable.
    • The receivers’ second report declined to provide updated valuations, precisely because doing so could prejudice the receivership.
    • In fact, the receivers were actively instructing agents to pursue private treaty sales.
    • In those circumstances, the mere existence of a notional “low end” auction figure did not require the trial judge to discard the expert’s going-concern valuation.
  • Questionable use of the extra £1.8 million loan:
    • The receivers noted that £530,000 of the £1.8 million had been used to redeem another loan to PEL.
    • However, “initial analysis” suggested that about £1.15 million may have been used by the husband to “divert value from PEL”, and further analysis was needed.
    • The remaining £120,000 serviced PEL’s existing facility.
    • This raised doubts about whether the additional indebtedness should properly be regarded as a genuine cost of preserving value, rather than a product of the husband’s manipulation.
  • Receivership caused by the husband’s own default:
    • The receivers had been appointed because the husband had engaged in egregious breaches of interim orders – failing to pay maintenance, legal services, and costs when ordered.
    • HHJ Hess had found that the husband had “absolutely no intention” of complying, that he would “resist all enforcement” and “manoeuvre his asset position to obstruct enforcement” whenever possible.
    • This conduct had also led to the breach of an injunction by increasing PEL’s borrowings without permission.
    • Consequently, the costs associated with receivership and any forced or sub-optimal sales were self-inflicted by the husband.
    • To reduce PEL’s valuation (thereby effectively diminishing the wife’s share) to reflect those costs would be to allow the husband to benefit from his own wrongdoing and, practically, to make the wife bear part of the very costs and arrears he owed her.

As to the alleged “inconsistency” with LR, the Court of Appeal drew a clear distinction:

  • In LR, the post-valuation events (completion and a much higher rent) clearly justified a more optimistic valuation, with the uplift moderated by deductions and the GDV “cushion”.
  • In PEL, the subsequent events were:
    • Primarily the consequence of the husband’s defiance of court orders; and
    • Insufficiently clear to justify any re-valuation away from the expert’s going-concern figure.

Accordingly, there was nothing irrational or legally flawed in the judge dealing with the two assets differently.

4.3 Appellate Restraint and Standards of Review

Paragraphs 70–74 of Moylan LJ’s judgment reaffirmed the modern approach to appellate review of first-instance decisions involving findings of fact, evaluative judgments, and exercises of discretion. Key authorities cited and applied include:

  • Henderson v Foxworth Investments Ltd [2014] 1 WLR 2600:
    • Lord Reed’s test: in the absence of an identifiable error of law, a critical factual finding with no evidential basis, a demonstrable misunderstanding or failure to consider relevant evidence, the appellate court will only interfere if the decision “cannot reasonably be explained or justified”.
  • Re R (Children) [2016] AC 76; [2015] UKSC 35:
    • Lord Reed stressed the limited function of an appellate court: where the lower court has applied the correct legal principles to the relevant facts, its evaluation is not open to challenge unless it was not reasonably open to it.
  • Lidl Great Britain Ltd v Tesco Stores Ltd [2025] 1 All ER 311:
    • Arnold LJ summarised that challenges to findings of fact require the finding to be “rationally insupportable” (per Volpi v Volpi), and challenges to multifactorial evaluations require a demonstrable error of law or principle.

Moylan LJ explicitly applied these standards. He also echoed the warning against “island hopping” in Fage v Chobani: an appellate court must beware of counsel seeking to undermine a fact-rich judgment by isolating individual points divorced from the broader context.

In Michael v Michael, the Court of Appeal held that:

  • The trial judge’s dealings with the tax liability, valuations of LR and PEL, and distribution of debts were all matters of fact-finding, evaluation, and discretion.
  • None of his conclusions were irrational, unsupported by evidence, or based on a misdirection of law.
  • Any alleged numerical inaccuracies were immaterial in the broader context of the award and did not render the outcome unjust.

4.4 New Points on Appeal and the Corporation Tax Act 2010

The husband attempted to raise a new line of argument, said to be grounded in the Corporation Tax Act 2010, as part of his attack on the lower-end tax scenario. This argument had not been advanced at trial and would have required fresh expert tax evidence.

Moylan LJ, relying on Singh v Dass [2019] EWCA Civ 360, reiterated the principle that new points which:

  • depend on expert evidence;
  • were not put to the trial judge; and
  • would or might have changed the course of the factual inquiries below;

are not normally admissible on appeal. The appellate court cannot fairly entertain such points without undermining the integrity of the trial process.

He therefore declined to consider this new CTA 2010 argument, and it played no part in the outcome.

5. Precedents and Authorities Cited: Their Role in the Judgment

The judgment integrates several leading authorities, using them to frame both the appellate approach and the scope of first-instance discretion.

5.1 Fage UK Ltd v Chobani UK Ltd [2014] EWCA Civ 5

  • Relied on by the husband to argue that the trial judge was required to make precise factual findings about the tax liability.
  • Countered by the Court of Appeal’s emphasis that:
    • the judge had fulfilled his fact-finding function by identifying the realistic range; and
    • he was not required to overstate the precision of his conclusions where the evidence did not permit it.
  • The case is also associated with the warning against “island hopping” – picking off individual elements of a carefully reasoned judgment on appeal.

5.2 Henderson v Foxworth Investments Ltd [2014] 1 WLR 2600

  • Cited for Lord Reed’s formulation of when an appellate court may interfere with findings of fact.
  • Used to underpin the high threshold for overturning the judge’s acceptance of the tax range and his asset valuations.

5.3 Re R (Children) [2015] UKSC 35; [2016] AC 76

  • Relied upon to reinforce the limited appellate role where the lower court has applied correct principles to the facts.
  • Supports the conclusion that Hess J’s dispositional choices, though contestable, were well within the range of reasonably open outcomes.

5.4 Lidl Great Britain Ltd v Tesco Stores Ltd [2025] 1 All ER 311

  • Arnold LJ’s summary of the standard of review for factual findings (“rationally insupportable”) and multifactorial evaluations (error of law or principle) is expressly adopted (para 73).
  • These standards are central to rejecting the husband’s attempts to re-open valuations and the treatment of tax.

5.5 Volpi v Volpi [2022] EWCA Civ 464

  • Noted through Lidl v Tesco, as the leading modern statement that appellate interference with fact-finding requires the decision to be rationally unsupportable.

5.6 Magmatic Ltd v PMS International Group plc [2016] UKSC 12; Actavis Group v ICOS Corp [2019] UKSC 15; Re Sprintroom Ltd [2019] EWCA Civ 932; Lifestyle Equities CV v Amazon UK Services Ltd [2024] UKSC 8

  • Cited collectively in Lidl v Tesco as authorities on appellate restraint in respect of multi-factorial evaluations (e.g. assessing similarity in trade mark cases or obviousness in patent law).
  • These authorities reinforce that assessments of value, risk, debt allocation and fairness in financial remedies are paradigmatic examples of such multi-factorial evaluations.

5.7 Singh v Dass [2019] EWCA Civ 360

  • Applied to refuse the husband’s attempt to introduce the new CTA 2010 point on appeal.
  • Confirms that appellate courts are generally unwilling to entertain new factual or expert-dependent arguments raised for the first time on appeal.

6. Clarifying Key Legal and Technical Concepts

6.1 Sham trust

A “sham trust” is one where the documents and formalities suggest a genuine trust structure, but the reality is different: the parties never intended the trustee to have the powers and duties set out in the trust instrument, and the assets are in substance controlled by someone else (here, the husband).

Once a trust is found to be a sham:

  • The court disregards its assumed separation of ownership.
  • Assets purportedly held “for beneficiaries” are treated as being beneficially owned by the person truly in control (here, the husband).
  • In tax terms, this recharacterisation can retrospectively alter the analysis of past transactions and trigger unexpected liabilities.

6.2 Reverse contingent lump sum

A reverse contingent lump sum is a financial remedy device whereby:

  • Party A (here, the wife) undertakes to pay Party B (the husband) a lump sum only if a specified future contingency occurs, such as:
    • An HMRC liability arising and being paid; or
    • A particular asset being sold or realised.
  • Until the contingency occurs, no payment is due.

This contrasts with ordinary contingent lump sums (which might be payable to a spouse on the occurrence of a future event, such as the sale of a property). The “reverse” element in Michael v Michael is that the payer is the economically stronger party (the wife) making a conditional contribution to a liability initially borne by the other party (the husband).

6.3 Matrimonial debt and unequal sharing

A “matrimonial debt” is typically any liability incurred during the marriage (and often in pursuit of matrimonial purposes) that the court takes into account when calculating the net matrimonial acquest.

Although the broad modern principle is one of sharing (particularly of matrimonial property), the court retains a wide discretion to:

  • Exclude certain debts from sharing (e.g. wasteful or reckless dissipation, or unilateral post-separation liabilities); or
  • Allocate debts unequally where justified by:
    • Conduct;
    • The source or purpose of the liability; or
    • Overall fairness under s 25 MCA 1973.

In Michael v Michael, the tax risk was plainly a matrimonial liability in timing and effect, but it arose out of the husband’s creation and defence of a sham structure. This justified imposing a greater share of the burden on him via the cap.

6.4 Minority discount

A “minority discount” is a reduction applied when valuing a shareholding that does not confer control (for example, a 50% or smaller stake in a company), to reflect the reality that minority interests often sell for less per share than controlling interests, because:

  • A minority shareholder cannot unilaterally direct management or force realisations;
  • There may be restricted markets for such interests; and
  • There may be shareholder agreements or practical constraints limiting exit.

In this case, the judge accepted the husband’s contention that a 20% discount should be applied to the values of companies in which H (UK) Ltd held 50% stakes, resulting in a deduction of about £3.8 million from the gross value.

6.5 Gross development value (GDV) and residual value

Gross development value (GDV) refers to the estimated market value of a completed development (e.g. a block of flats or offices) once fully built and let/sold.

Residual value is the value of a development site or incomplete scheme today, calculated by deducting from the GDV all costs still to be incurred (construction, finance, professional fees, contingencies) and profit margins. It represents the price a rational developer would pay now to undertake the future development work.

In LR’s case:

  • GDV: £32.5 million (expert’s view).
  • Residual value (incomplete): £15.36 million.
  • The wife’s challenge was that it was unrealistic to treat LR as if it were to be sold incomplete, given the near-completion and subsequent letting.

6.6 Receivership

Receivership in this context involved the court appointing receivers over PEL’s property portfolio to enforce sums the husband owed to the wife. The receivers:

  • Step into control of specified assets;
  • May manage or sell properties;
  • Use sale proceeds to discharge secured debts and then, as directed, pay amounts towards the husband’s liabilities to the wife.

Receivership often entails additional costs and can lead to distressed sales. Here, these consequences were traced directly to the husband’s persistent non-compliance with court orders.

7. Impact and Future Significance

7.1 Managing uncertain tax liabilities in financial remedies

Michael v Michael provides valuable guidance for cases where a financial remedy court is confronted with a large, speculative or contingent tax liability, especially arising from:

  • Sham transactions or structures;
  • Retrospective re-characterisation of past dealings (e.g. as income distributions); or
  • Pending HMRC enquiries with wide discretionary outcomes.

Key points for practitioners and judges:

  • The court is not obliged to pick a single number if the expert evidence only supports a wide range. It may:
    • Adopt that range as its factual finding; and
    • Use contingent mechanisms to allocate risk fairly between the parties.
  • Reverse contingent lump sums are endorsed as an appropriate and flexible tool for dealing with uncertain liabilities like HMRC assessments.
  • The court can legitimately cap one party’s exposure to a shared liability where:
    • The liability arises largely from the other party’s conduct (e.g. sham trusts, aggressive tax schemes, concealment); and
    • Overall fairness (including needs, resources, illiquidity and risk) justifies such a cap.

7.2 Unequal sharing of matrimonial debts based on responsibility

The case confirms that even where a liability is properly categorised as a matrimonial debt, the court may depart from 50:50 sharing where one party is plainly more responsible for its incurrence. In particular:

  • Tax arising from a sham arrangement controlled by one spouse may justifiably be charged disproportionately to that spouse.
  • This will often be framed through the conduct factor (s 25(2)(g) MCA 1973), but more broadly it reflects the overarching inquiry into fairness.

This has practical implications for spouses who design and implement aggressive or opaque tax structures: if those structures collapse or are re-characterised, they may not be able to insist that their former spouse bear half the fallout.

7.3 Post-valuation events and business valuations

The decision reinforces that:

  • The court has a discretion to take account of post-valuation events (such as grant of planning permission, completion, improved letting) where fairness demands, but is not compelled to do so in every case.
  • It is legitimate to treat different assets differently, depending on:
    • the nature and reliability of the post-valuation evidence;
    • whether the change in value is truly crystallised or still speculative; and
    • whether the change results from a party’s own default (as with receivership triggered by non-compliance).

Challenges on appeal to such selective treatment will face a steep uphill battle, given the evaluative nature of the exercise and the high threshold for appellate interference.

7.4 No benefit from one’s own default

The court’s robust stance in relation to PEL sends a strong message:

  • A party who persistently defies court orders and thereby triggers receivership and enforcement costs cannot expect the court to discount asset values to reflect the very costs caused by that behaviour.
  • To do so would risk making the innocent party bear a portion of their own unpaid costs and arrears, and would undermine the authority of interim orders.

This principle may be invoked in future cases where a non-compliant party seeks to argue that enforcement action has diminished the asset pool and should therefore reduce the other party’s award.

7.5 Appellate discipline in financial remedy cases

Finally, Michael v Michael contributes to the wider jurisprudence on appellate restraint in family finance cases:

  • Relatively small valuation differences (4% of the total estate here) will rarely justify an appeal.
  • The Court of Appeal will be wary of attempts to challenge discrete aspects of a complex, fact-sensitive judgment through “island hopping”.
  • New legal or factual points that would have required evidence at trial will generally be excluded on appeal.

For practitioners, the message is clear: appeals in financial remedies are not re-hearings. Unless a genuine error of law or a plainly unsustainable evaluation can be shown, the trial judge’s approach will stand.

8. Conclusion

Michael v Michael [2025] EWCA Civ 1668 is a significant authority in three overlapping areas of family financial law:

  1. Uncertain tax liabilities and sham structures – The Court of Appeal endorses the use of reverse contingent lump sums and caps to manage large, unpredictable tax exposures. A trial judge is not compelled to select a single figure where the evidence supports only a wide range; instead, the court may design a contingent mechanism to allocate risk fairly.
  2. Allocation of matrimonial debts – Even where a liability is matricial in character, the court can justifiably place a greater burden on the party primarily responsible for creating or maintaining the structure giving rise to that liability, especially where their conduct has been criticised.
  3. Valuation and appellate restraint – The decision confirms that selective consideration of post-valuation events, and the rejection of attempts to revalue assets downward where such changes result from a party’s own non-compliance, fall squarely within the trial judge’s discretion. Appeals on such issues face a high threshold, particularly where the overall impact is modest in percentage terms.

Overall, the judgment underscores the centrality of fairness, the importance of conduct and responsibility in allocating complex liabilities, and the need for finality and deference to first-instance fact-finding in financial remedy litigation. It provides a thoughtful and pragmatic template for dealing with contingent tax risks in sophisticated, trust- and company-heavy financial remedy cases.

Case Details

Year: 2025
Court: England and Wales Court of Appeal (Civil Division)

Comments